External Two Stage DEA Analysis of Bank Efficiency in West Balkan Countries

Abstract

Since the beginning of the application of the Data Envelopment Analysis (DEA) model in various areas of the economy, it has found its wide application in the field of finance, more specifically banks, in the last few years. The focus of this research was to determine the sustainability of the intermediate function of banks, especially in recent years when interest rates on deposits have been at a minimum level. The research was divided into two parts, wherein the first part determined the efficiency of the intermediate function of banks in the countries of the Western Balkans in the period from 2015 to 2019. The second part approached the regression analysis in which we determined the influence of the bank size, type of bank, and mergers and acquisitions (M&A) activity on the defined efficiency. In the first stage we applied the output-oriented DEA model using deposits, labor costs, and capital as input variables; on the other side, we used loans and investments as output variables. We used data from the revised financial statements of the banks operating in Serbia, Bosnia and Herzegovina, Montenegro, North Macedonia, and Albania. The results of our study showed that there is a difference in efficiency levels between countries and within countries in the considered time period. Furthermore, Tobit regression analysis showed a significant and negative influence of the bank type and M&A on relative technical efficiency of banks, and a positive and significant relationship between bank size and relative efficiency. These findings suggest that large commercial banks can sustain on the West Balkan market. It is to be expected that less efficient small banks will be taken over by large and more efficient banks.

Do business strategies vary across firms in the banking industry? New perspectives from the bank size–profitability nexus

“…It is obvious that as banks grow in size, they gain cost‐efficiency (Anagnostopoulos et al, 2022; Halkos & Tzeremes, 2007). This translates into declines in average costs, and it realizes the economies of scale and scope (Adusei, 2016; Berger et al , 1993; Milenković et al , 2022; O'Connell, 2022; Rehman et al , 2022; Yin et al , 2013). This finding supports the research outcome of Anagnostopoulos et al (2022) and Ding and Sickles (2018), which stresses that banks enjoy cost‐efficiency as they grow in size but negates the research outcomes of Rakshit and Bardhan (2022b) establishes that growth in bank size leads to bank inefficiency.…”

“…Their conclusion is premised on the fact that larger banks enjoy economies of scale and thus can increase their cost-efficiency over the smaller banks (Milenkovi c et al, 2022). Conversely, a strand of studies documents the negative effect of bank size and profitability (Amato & Burson, 2007;Banchuenvijit, 2012;Singh et al, 2021) and bank size on cost-efficiency (Ding & Sickles, 2018;Fang et al, 2019;Hadhek et al, 2018).…”

“…Similarly, focusing on cost‐efficiency as a measure of performance, some studies discover a significant positive relationship between bank size (log of the total asset) and efficiency (Anwar, 2018; Eriki & Osifo, 2015; Karray & Chichti, 2013; Moyo, 2018; Oluitan et al, 2015; Otero et al, 2020). Their conclusion is premised on the fact that larger banks enjoy economies of scale and thus can increase their cost‐efficiency over the smaller banks (Milenković et al , 2022). Conversely, a strand of studies documents the negative effect of bank size and profitability (Amato & Burson, 2007; Banchuenvijit, 2012; Singh et al, 2021) and bank size on cost‐efficiency (Ding & Sickles, 2018; Fang et al, 2019; Hadhek et al , 2018).…”

Investments in Tourist Destinations for Increasing Efficiency in the Future Serbian Tourism Sector

“…Parametric methods focus on average tendencies and deviations from it, while nonparametric methodology focuses on the extreme observations (Cvetkoska & Fotova-Čiković, 2021). One of the most popular nonparametric methodology is DEA with various models and modification that can be applied for efficiency assessment in many different fields: from banking (Milenković et al, 2022;Nourani et al 2021;Bod'a & Zimková, 2015), insurance (Radovanov et al, 2021) Authors Ilić and Petrevska (2018) used DEA for efficiency assessment of tourism in Serbia and surrounding countries. DEA model had two input variables (tourism expenses and number of beds) and three output variables (number of arrivals, the number of nights spent and tourism revenue).…”

DEA efficiency of Serbian banks: Comparison of three approaches

“…External two-stage models use the combination of DEA analysis in the first stage and a regression analysis in the second stage (Milenković et al, 2022;Paleckova, 2019;Simar & Wilson, 2011;Sufian, 2010). In these models, DEA efficiency index is measured first, and some regression methods like Bootstrap, Tobit, OLS, AHP, ANN and others are used in the second stage (Henriques et al, 2020).…”